Canada

Inspired by notions of intrinsic motivation (Schmidhuber, 2010) and play as proposing and solving arbitrary problems (Chu and Schulz, 2020b), we report initial progress toward computational modeling of playful goal generation. We create an embodied, 3D environment resembling a child&#39;s room, and ask study participants to play in the environment and then create a scorable game. We propose to model games using a domain-specific language, which represents each game as a computer program. These programs act as reward-generating functions, mapping states visited by an agent as they play a game to the score they should receive. We then analyze our corpus of program representations to highlight four key aspects of human games that would contribute to constructing effective computational models of game generation: creativity, compositionality, common sense, and context sensitivity.

CogSci 2022

Creativity, Compositionality, and Common Sense in Human Goal Generation

Inspired by notions of intrinsic motivation (Schmidhuber, 2010) and play as proposing and solving arbitrary problems (Chu and Schulz, 2020b), we report initial progress toward computational modeling of playful goal generation. We create an embodied, 3D environment resembling a child's room, and ask study participants to play in the environment and then create a scorable game. We propose to model games using a domain-specific language, which represents each game as a computer program. These programs act as reward-generating functions, mapping states visited by an agent as they play a game to the score they should receive. We then analyze our corpus of program representations to highlight four key aspects of human games that would contribute to constructing effective computational models of game generation: creativity, compositionality, common sense, and context sensitivity.

poster

Join us for CogSci2022! Our 44th Annual Conference will be centered around the theme of diversity, with an accompanying range of keynotes and invited speakers who themselves highlight the diversity of our field. Our three symposia are Cognition across Diverse Populations, Cognition within Diverse Societies, and Global Challenges for Cognitive Science. The conference is being offered in a fully hybrid environment for onsite and online attendees.
 
 [**The Program at a Glance can be found here**](https://cognitivesciencesociety.org/wp-content/uploads/2022/07/Cogsci-2022-PAAG-Detailed-July-12final-1.pdf)
 [**The Detailed Program can be found here**](https://cognitivesciencesociety.org/wp-content/uploads/2022/08/DeArmond-CogSci2022-program-v113-final.pdf)

This conference requires a registration to attend. Please use the link below to secure your ticket. The website will be available to all registered attendees on July 25th. 

[Register to Attend](https://cognitivesciencesociety.org/registration/)

Registration Is Required

The 44th Annual Meeting of the Cognitive Science Society highlighted research on the theme Cognitive Diversity in addition to the full breadth and diversity of research topics offered by the society’s membership. 





Many models in Cognitive Science require data to calibrate parameters. Some modelers calibrate their models’ parameters for each individual in a data set, and others work at the aggregate level. Generally, the accuracy of a model is judged by the degree to which human data are replicated, and the model parameters are interpreted accordingly. It is not too surprising that models that are developed for a particular task and fit to each individual’s data in such a task replicate the human data well. The question is, however, whether those models can make predictions in the absence of human data. In this paper, we present a theory-driven model of a well-known sequential decision task (the Balloon Analog Risk Task, BART) which is able to make predictions in the absence of human data. The cognitive model is grounded on the processes and mechanisms of Instance-Based Learning (IBL) Theory of experiential choice. We demonstrate the simulation predictions from an IBL model and those of a well-known model of the BART, which depends on the fits to human data. We further show that when making predictions without data, the IBL model provides predictions that are both theoretically founded and accurate, while the Two-Parameter model performs much worse than when fit to data. We conclude with a discussion of the benefits of making theory-based predictions in the absence of human data for our community.

Making Predictions Without Data: How an Instance-Based Learning Model Predicts Sequential Decisions in the Balloon Analog Risk Task

Children are navigating a world with massive amounts of auditory input, sometimes relevant while other times purely noise, and must somehow make sense of it all. The early auditory environment is critical for speech perception and recognition, auditory discrimination, and word learning, all of which support language outcomes. What strategies do children use to learn in noisy environments? One potential strategy is environmental selection, which allows children to seek environments that align with particular goals. In the current paper, we examined whether children and adults make decisions about their environments by integrating auditory information and goal-states. While 3- and 4-year olds struggle with discriminating the level of noise in noisy speech streams (and likely do not use this information for environmental selection), 5-year-old children and adults can. Further, we show initial evidence that they can use this information to reason about acoustic environments that are consistent with specific goals.

Selection of goal-consistent acoustic environments by adults and preschool-aged children

Though high-confidence identifications from a police lineup can provide strong evidence of guilt, confidence in a lineup rejection often provides little additional information beyond the rejection itself. Here, we investigate whether participants use different decision rules for positive identifications vs. lineup rejections. In this experiment, we compared confidence ratings from a standard simultaneous lineup (where confidence in the lineup rejection is not applied to any particular face) to the confidence ratings from a rate-them-all simultaneous lineup (where confidence that a person is or is not the perpetrator is applied to every face in the lineup). We hypothesize that, in the case of a lineup rejection, confidence in the standard condition will be best predicted by the average memory signal in the rate-them-all condition. The possible reliance on an average memory signal could help to explain why previous research has not found a strong confidence-accuracy relationship for lineup rejections.

Decision Variables in the Case of Police Lineup Rejections

From writing to hiking, people’s real-world sequential decision-making often benefits from “undoing” (e.g. deleting sentences or backtracking). Surprisingly, undoing has not been studied in experiments on human planning. To investigate how much, when, and why people undo, we introduce a task that is a cross between the “Traveling Salesperson” and the “Knapsack” problems with an undo option. Within a length budget, subjects sequentially connect as many dots as possible on a map. On “undo” trials, they are allowed to take back actions without constraints. We find that undoing is beneficial, that subjects exhibit great individual variability in the number of undos, that undos are more frequent after errors than after correct actions, and that long response times tend to precede sequences of undos. Together, these results suggest that undo actions serve a dual role of correcting errors and of exploring alternative paths, where path evaluation benefits from full play-outs.

Undoing in human planning

Geometric analogy problems remain an intriguing part of intelligence scales, which is closely correlated to many cognitive studies, such as perception, conception, memory, abstract and inductive reasoning. The problems not only target the most fundamental element --- analogy-making --- in human cognition, but also require integration of multiple components and stages: looking at the test booklet, thinking for a minute or two, and deciding the answer. Great efforts and achievements have been made to explain different individual aspects of this process. In this paper, we take a more holistic approach from the perspective of problem-solving, by modeling the entire process, from the moment the visual stimuli are received to the moment an answer is decided. Therefore, we explore how the final solution can be built upon visual inputs and necessary components that lie between the perceptual input and conceptual output. Particularly, we designed a novel similarity metric and a correspondence-finding method based on mapping and optimization. With these two basic blocks, we implemented a computational model, and report our initial results on a classical problem set.

An End-to-End Imagery-Based Modeling of Solving Geometric Analogy Problems

In STEM instruction, the effectiveness of teaching by analogy is often limited by students’ focus on superficial features of the source and target exemplars; the strategy of progressive alignment (moving from perceptually similar to different targets) has been suggested to address this issue (Gentner & Hoyos, 2017). In contrast, computational models suggest maximizing surface feature variation to improve the relation learning, and human behavioral studies find a relational bias in learning (Austerweil & Ehrens, 2018). Here, participants were explicitly instructed to match stimuli based on relations while perceptual similarity of stimuli varied parametrically. We found that lower perceptual similarity reduced accurate relational matching (F = 9.53, p<.001), and observed a similar trend for reaction times. This finding demonstrates that perceptual similarity may interfere with relational judgements, but also hints at why progressive alignment may be effective. Implications for instructional sequence design will be discussed.

Perceptual Similarity Affects Relational Judgements

Understanding abstract relations, and reasoning about various instantiations of the same relation, is an important marker in human cognition. Here we focus on development of understanding for the concept of antonymy. We examined whether four- and five-year-olds (N= 67) are able to complete an analogy task involving antonyms, whether language cues facilitate children’s ability to reason about the antonym relation, and how their performance compares with that of two vector-based computational models. We found that explicit relation labels in the form of a relation phrase (“opposites”) improved performance on the task for five-year-olds but not four-year-olds. Five-year-old (but not four-year-old) children were more accurate for adjective and verb antonyms than for noun antonyms. Two computational models showed substantial variability in performance across different lexical classes, and in some cases fell short of children’s accuracy levels. These results suggest that young children acquire a solid understanding of the abstract relation of opposites, and can generalize it to various instantiations across different lexical classes. These developmental results challenge relation models based on vector semantics, and highlight the importance of examining performance across different parts of speech.

Children’s Acquisition of the Concept of Antonym Across Different Lexical Classes

Analogy plays an important role in cognitive development, but children often need cognitive supports to draw correct ones. Here, we examined the role of conceptual knowledge in proportional analogies, which are often depicted as a simple exercise in pattern completion. In Study 1, adults and children (N = 321) completed 4-term analogy tasks featuring letters, lines, integers, or fractions. Performance was lowest for fractions, and strongly impacted by educational background. In Study 2, we conducted an educational intervention focusing on either conceptual knowledge, procedural knowledge, or both for 3rd-to-5th graders (N = 343) using a pretest-training-posttest design. Children with poor pretest magnitude knowledge were more likely to fail analogical reasoning, and training on conceptual knowledge that fractions denote magnitudes improved children’s analogies. Together, these studies indicate that knowledge of fractional magnitudes is important to proportional analogy.

Conceptual Prerequisites for Proportional Analogy

One of the advantages of using metaphorical expressions over literal ones might be that speakers can convey not only the intended property, but also its degree. 
For example, when hearing “John is a shark”, the listener might infer that the speaker aims to communicate that John is as mean as a typical shark. 
We present experimental findings supporting this hypothesis, along with a novel metaphor interpretation model, which is implemented within the Rational Speech Act framework.
We compare our model's predictions to those of an existing RSA model of metaphor understanding, within which the listener infers just the presence or absence of a feature as opposed to its degree, and find that our model produces a significantly better fit.

Pragmatics of Metaphor Revisited: Modeling the Role of Degree and Salience in Metaphor Understanding

Structural priming is a heavily studied and multi-faceted phenomenon. Essentially, when other factors are equal, syntax shows a tendency to repeat across utterances with potential gains in fluency (Pickering & Ferreira, 2008). Lexical repetition, particularly repetition of the verb between utterances enhances structural priming significantly (Mahowald et al., 2016), but this lexical-boost is not as long-lasting (Hartsuiker et al., 2008). Previous research on structural priming has heavily leaned on measures of syntactic choice, but a few studies have also measured initiation time (Smith & Wheeldon, 2001). The current study looks to differentiate effects of lexical and structural repetition benefits in initiation latency as a measure of potential fluency increases in sentence production. Results suggest that benefits in fluency are dependent on lexical and structural repetition.

Premium content

Creativity, Compositionality, and Common Sense in Human Goal Generation

Downloads

Next from CogSci 2022

Making Predictions Without Data: How an Instance-Based Learning Model Predicts Sequential Decisions in the Balloon Analog Risk Task

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES